The present invention relates to a device for measuring the temperature of a melt, particularly of a molten metal, for example molten steel, with an optical fiber.
The Electric Arc Furnace (EAF) process for the production of molten steel is a batch process made up of the following operations: furnace charging of metallic components, melting, refining, de-slagging, tapping and furnace turnaround. Each batch of steel, called a heat, is removed from the melting furnace in a process called tapping and, hence, a reference to the cyclic batch rate of steel production is commonly a unit of time termed the tap-to-tap time. A modern EAF operation aims for a tap-to-tap cycle of less than 60 minutes and is more on the order of 35-40 minutes.
Many of the advances made in EAF productivity that promote rapid tap-to-tap times possible are related to increased electrical power input (e.g., in the range of 350-400 kWh/t), and alternative forms of energy input (e.g., oxygen lancing, oxy-fuel burners) into the furnace. Most advanced EAF operations consume on the order of 18-27 Nm3/t of supplemental oxygen which supplies 20-32% of the total power input. In addition, improvements to components which allow for faster furnace movement have reduced the amount of time in which the furnace stands idle. The industrial objective of EAF operators has been to maximize the furnace power-on time, resulting in maximum productivity in order to reduce fixed costs, while at the same time gaining the maximum benefit from the electrical power input. The majority of time consumed in the production of one heat of steel in the EAF is in the process step of melting.
The melting period is the heart of EAF operations and, in the majority of modern EAFs, occurs in a two stage process. Electrical energy is supplied via graphite electrodes and is the largest contributor in the melting operation. To melt steel scrap, it takes a theoretical minimum of 300 kWh/t. To provide the molten metal with a temperature above that of the melting point of steel requires additional energy. For typical tap temperature requirements, the total theoretical energy required usually lies in the range of 350-400 kWh/t. However, EAF steelmaking is only 55-65% energy efficient and, as a result, the total equivalent energy input is usually in the range of 650 kWh/t for most modern operations with 60-65% supplied by electric power, the remaining requirements supplied by fossil fuel combustion and the chemical oxidation energy of the refining process.
During the first metallic charge, an intermediate voltage tap is usually selected until the electrodes can sufficiently bore into the scrap. The position of unmelted scrap between the electrode arc and the side wall of the melting vessel protects the furnace structure from damage such that a long arc (high voltage) tap can be used after boring. Approximately 15% of the scrap is melted during the initial bore-in period. Fossil fuel combustion added through special nozzles in the furnace wall contributes to scrap heating and thermal uniformity. As the furnace atmosphere heats up, the arcing tends to stabilize and the average power input can be increased. The long arc maximizes the transfer of power to the scrap and the beginnings of a liquid pool of metal will form in the furnace hearth. For some specific EAF types, it is a preferable practice to start the batch melting process with a small pool held over from the previous heat called a “hot-heel”.
When enough scrap has been melted to accommodate the volume of second charge, the charging process is repeated. Once a molten pool of steel is generated in the furnace, chemical energy may now be supplied via several sources, such as oxy-fuel burners and oxygen lancing. Oxygen can be lanced directly into the bath once the molten metal height is sufficient and clear of obstructive scrap.
Nearing the time that the final scrap charge is fully melted, the furnace sidewalls can be exposed to high radiation from the arc. As a result, the voltage must be reduced or the creation of a foamy slag that envelops the electrodes. The slag layer may have a thickness of more than a meter while foaming. The arc is now buried and will protect the furnace shell. In addition, a greater amount of energy will be retained in the slag and is transferred to the bath resulting in greater energy efficiency. This process will create a lot of heat in the slag layer covering the steel, resulting in temperatures that are up to 200° C. higher than the steel temperature creating a very unique and difficult environment for process control measurements for reasons explained later.
Reducing the tap-to-tap time for a heat, in many instances and especially in modern EAF operations operating with a hot heel, oxygen may be blown into the bath throughout the heat cycle. This oxygen will react with several components in the bath including aluminum, silicon, manganese, phosphorus, carbon and iron. All of these reactions are exothermic (i.e., they generate heat) and will supply energy to aid in the melting of the scrap. The metallic oxides which are formed will eventually reside in the slag.
When the final scrap charge and raw materials are substantially melted, flat bath conditions are reached. At this point, a bath temperature and a chemical analysis sample will be taken to determine an approximate oxygen refining period and a calculation of the remaining power-on time until tap.
Regardless of the specific local processing steps that may vary depending upon the utilization of available raw material, furnace design, local operating practices and the local economies of production, it is evident that many forms of energy inputs to the furnace may be employed in a variety of strategies in order to minimize the tap-to-tap time and improve energy efficiency during the conversion of solid scrap and slag components to molten steel and slag at the correct chemical composition and desired temperature for tapping.
As in other steelmaking processes, the tap-to-tap production process of an EAF is guided by mathematical models that take into account the quantity and quality of raw materials in order to predict the process end point given the energy inputs and heat outputs. A listing of such variables can be found in EP 0747492 A1. Many of the process models used to control and predict EAF performance are well known in the art. When compared to the classic steelmaking process of blast furnace to converter, the variance of the raw materials used in the EAF process is much higher and as such require constant adjustments. One of several information inputs to these models required to correct and guide the process is the molten metal temperature.
Providing the EAF operator with the best and most recent molten metal temperature information should satisfy the following requirements:
an accurate temperature representative of the bulk metal,
fixed immersion depth independent of the furnace tilt,
continuously or nearly continuously available, and
bath level determination for immersion depth adjustments.
Typically, a temperature measurement of the molten metal is accomplished using well known disposable thermocouples such as described in U.S. Pat. No. 2,993,944. Such thermocouples can be immersed manually by an operator with a steel pole with adapted electrical wiring and connections to convey the thermocouple signal to appropriate instrumentation. Additionally, many automatic thermocouple immersion mechanical systems are now utilized to provide thermocouple immersions, such as those publically available from www.more-oxy.com or described in the literature Metzen et al., MPT International April 2000, pp. 84.
Once pooling of molten metal is established, the bath temperature will slowly increase. The higher the content of the non-molten scrap the lower the rate of temperature increase will be for a given energy input. Once all the scrap is molten, the temperature of the bath will increase very rapidly, in the order of 35° C.-70° C./minute toward the end of the process. In order to predict the optimum process end, the time that the metal is ready to tap, the process control models need to have temperature information that is accurate and at a sufficiently high frequency of measurements to create an accurate forecast of the best moment to stop the various energy inputs. The measuring process using robotic immersion devices requires that an access hatch, typically the slag door, a general description of which appears in U.S. Patent Application Publication No. 2011/0038391 and in U.S. Pat. No. 7,767,137, is opened to allow insertion of a mechanical arm supporting a disposable thermocouple. In most modern operations, this door is also used to provide access to the furnace for oxy-fuel burners and oxygen lances that are brought into position with a similar manipulator to that of the immersion lance. More recently, several additional ports may also be provided around the circumference of the furnace shell for burners as described in U.S. Pat. No. 6,749,661.
Opening of the slag door for the purpose of obtaining temperature late in the process allows a large amount of air to enter the furnace. Consequences of this opening are cooling the local area and providing a source for nitrogen. During arcing, nitrogen is converted to NOx which is an undesirable effluent of the EAF process. While it is necessary to deslag the furnace through this opening, the use of robotic immersion equipment also utilizing this opening to take temperatures exposes the furnace interior to unnecessary nitrogen ingress and unintentional de-slagging during periods when repeated temperature measurements are required.
With a rapid temperature rise during the end stages of the metal refining process, the update time for a process control model under the best of circumstances cannot keep up with modern high powered furnaces. Ideally, rapid temperature updates during the end of refining and continuous temperature information during the last minutes prior to tap provide the best combination for model prediction accuracy and end point determination. A realistic test-to-test time of one minute for typical robotic systems limits the usefulness of spot measurements of such a dynamic process. Conventional disposable thermocouples and robotic immersion equipment suffer from several additional limitations besides a low sampling frequency that ultimately reduces the predictive success of the process models when used for accurate end point decisions.
During the melting and refining processes, the bath will have a temperature gradient whereas the surface of the bath will have a significantly higher temperature than that of the bulk molten metal. Hot and cold spots of metal are located throughout the furnace interior necessitating the use of specialized burners and directional fossil fueled heaters to help homogenize the interior. As indicated in EP 1857760 A1, one cold spot is in the area of the slag door where the immersion of disposable thermocouples typically occurs due to the large access requirements of the typical robotic immersion equipment. An EAF has the ability to “rock” furnace, that is, to tilt the horizontal position of the furnace, front to back, in order to further homogenize the bath, deslag and tap the furnace, as described in U.S. Pat. No. 2,886,617.
Most all robotic immersion devices are mounted in the area of the slag door and are mounted on the operating floor, and thus do not tilt through the angle of the tilted furnace. Consequently, such manipulators cannot position disposable thermocouples into the bath at all times and under all circumstances. Furthermore, the immersion depth of a thermocouple is linked to the articulation of the mechanical arm of the robotic device and, as such, cannot readily adjust to a bath level change due to the angle of the furnace tilt. While it is important to repeatedly measure in a location that reflects the bulk temperature for the purpose of the operating models of the EAF process, the actual temperature measurements taken with either a manual or automatic lance show difficulties towards stable immersion depths, not available while the position of the immersion lance is not aligned to the rocking of the furnace and the actual bath level, and not in a location conducive to temperature accuracy.
There are numerous temperature measuring devices in the prior art installed in a variety of steelmaking vessels that utilize permanent optical light guides to focus the radiation toward the optical detectors. Examples of such prior art devices can be found in JP-A 61-91529, JP-A-62-52423, U.S. Pat. No. 4,468,771, U.S. Pat. No. 5,064,295, U.S. Pat. No. 6,172,367, U.S. Pat. No. 6,923,573, WO 98/46971 A1 and WO 02/48661 A1. The commonality of this prior art is that the optical guides are permanent and, as a result, need to be protected from damage using complicated installations. These protective means may comprise gas purging to either cool the assembly or remove the metal from physical contact with the optical element, layers of protective sheathing that are relatively permanent or slightly erodible with the lining of the steelmaking vessel and complicated emissivity correction of the light wavelength(s) and intensity in order to determine an accurate temperature.
JP-A-08-15040 describes a method that feeds a consumable optical fiber into liquid metal. The consumable optical fiber, such as disclosed in JP-A-62-19727, when immersed into a molten metal at a predictable depth receives the radiation light emitted from the molten metal at blackbody conditions, such that the intensity of the radiation using a photo-electric conversion element mounted on the opposite end of the immersed consumable optical fiber can be used to determine the temperature of the molten metal. The scientific principle of the prior art concisely detailed in P. Clymans, “Applications of an immersion-type optical fiber pyrometer”, is that the optical fiber must be immersed at a depth to achieve blackbody conditions. Continuous measurements of molten metals using consumable optical fiber and equipment necessary to feed long lengths of coiled material to a predetermined depth are well known in the art, such as EP 0806640 A2 and JP-B-3267122. In harsh industrial environments where the consumable optical fiber is immersed into higher temperature metals or in the presence of metals with a slag covering maintaining a predetermined depth during the period of time the measurement should take place has proven to be difficult due to the inherent weakness in the optical fiber as its temperature increases. It has become necessary to protect the already metal covered fiber with additional protection such as gas cooling as disclosed in JP-A-2000-186961, additional composite materials layered over the metal covered fiber as disclosed in EP 655613 A1, insulating covering as disclosed in JP-A-06-058816, or additional metal covers as disclosed in U.S. Pat. No. 5,163,321 and JP-B-3351120.
The above improvements for high temperature use have the disadvantage of dramatically increasing the cost of the consumable fiber assembly in order to provide a continuous temperature reading. Although not exactly identical to the conditions encountered when measuring higher temperatures in an EAF, JP-B-3351120 is useful to have an appreciation of the speed of consumption of the optical fiber. In the disclosed example using a very complex mechanical device for feeding, an optical fiber from a coil is used. The coil consists of the metal covered optical fiber covered again with additional 3 mm thick stainless steel tubing. The disclosed calculations recommended for improved temperature accuracy for continuous temperature measurements in iron of a blast furnace tap stream is an astonishing 500 mm/s. The cost of the optical fiber and its enclosing stainless steel outer tube are costly to consume at this recommended feeding rate.
A practical economy of continuous temperature measurements depends upon consuming the least amount of fiber possible while still obtaining the benefit of continuous information. Bringing the optical fiber to the measuring point with the least amount of exposed fiber is described in U.S. Pat. No. 5,585,914 and JP-A-2000-186961, where a single metal covered fiber is fed through a permanent nozzle that could be mounted in the furnace wall and through which gas is injected. While these devices can successfully deliver the fiber to the measuring point, they become a liability due to clogging and continuing maintenance. At stages in the feeding mode, vibration is required to prevent the fiber from welding to the nozzle. If the port is blocked or closes due to inadequate gas pressure, the measurement is terminated with no possibility of recovering until the nozzle is repaired. EP 0802401 A1 addresses the problem of a blocked opening to the furnace with a series of punch rods and guide tubes positioned on a movable carriage, providing a tool set for addressing whichever problem prevents the fiber from passing through the nozzle. However, these are strategies to unblock a closed access port from which no measuring data can be obtained. Once these ports are blocked there is no possibly to obtain temperature data, which could be at critical times in the steelmaking process.
An additional problem arises for continuously fed optical fibers that further increase the cost of measurement and the complexity of the immersion equipment. The immersion type optical fiber only maintains its optical quality, and thus returns and accurate temperature if it stays protected against heat and contamination or is renewed at a rate higher than its degradation rate. The optical signal of the bath temperature is accurately obtained in blackbody conditions for the part immersed in the molten steel. However, the remaining un-immersed portion part above must remain a perfect light guide. At elevated temperatures, devitrification of the optical fiber will occur, the transmissivity of the light decreases and an error in temperature as a function of decreased intensity increases. JP-A-09-304185 and U.S. Pat. No. 7,891,867 disclose a feeding rate method where the speed of fiber consumption must be greater than the rate of devitrification, thereby assuring that a fresh optical fiber surface is always available. Simple laboratory testing shows that the optical signal stays stable during a very short period, being around 1.0 s at temperatures below 1580° C. and only 0.1 s while immersed at 1700° C. Although a solution for lower temperature metals, the speed of feeding optical fiber at a speed greater than the devitrification rate for elevated temperature testing is expensive for a simple metal covered optical, fiber. In the case of measuring elevated temperatures in the harsh conditions of an EAF, the prior art disclosed extra protection methods are also consumed at the same rate of as the optical fiber. This becomes prohibitedly expensive for the above mentioned double covered optical fibers.
JP-A-2010-071666 discloses a fiber optical temperature measuring device for measurements in molten metal using an airtight environment and a measuring lance having an airtight sealing between lance tube and optical fiber.